RAID combines several disks into one array so they go faster, survive a drive failure, or both. Here is what it means in plain English — then size any level in the RAID calculator.
The basic idea
RAID (redundant array of independent disks) presents several physical drives to your system as a single logical volume. A RAID controller (a hardware card or software) decides how data is spread across the drives, using three techniques in different combinations: striping for speed, mirroring for safety, and parity for space-efficient redundancy.
The point is resilience and/or performance beyond a single disk. Most RAID levels let the array keep running — and rebuild — after a drive dies, so a hardware failure doesn't mean downtime or data loss.
The three techniques
Striping (RAID 0) splits data across drives for speed but adds no protection. Mirroring (RAID 1) keeps a full copy on another drive. Parity (RAID 5/6) stores a checksum that can rebuild a missing drive without keeping a whole second copy — more space-efficient than mirroring.
Real-world levels mix these. RAID 10 stripes across mirrors; RAID 5/6 stripe with one or two parity blocks; ZFS RAIDZ does parity in software. The RAID levels guide covers each, and the calculator shows the capacity, resilience and speed trade-off.
What RAID doesn't do
RAID protects against drive failure — not against deletion, corruption, ransomware or site loss, all of which it replicates across every disk. RAID is availability, not backup; you still need a separate, ideally air-gapped, copy. See RAID is not a backup.
It also isn't free capacity: every redundant level gives up some space (and parity levels add a write penalty). The calculator makes those costs explicit so you choose with eyes open.